P a r t IV
Part IV provides several typical tutorials sessions in order to give practical information to users of the Turbo-Frodo program. These sessions include, for example, a general users' tutorial, which describes how to display and model a protein, a NMR users' tutorial, which shows how to display, correct, and refine a calculated structure, or a Non-Crystallographic Symmetry tutorial, which explains how to define a symmetry operation and display NCS-related molecules and maps, among others.
CHAPTER 16
General Users' Tutorial
This chapter describes how a Turbo-Frodo session for displaying and modeling a protein may be run.
The Heap File
Heap File is the data bank in which all your molecules are stored. It is an infinite file and is limited only by your disk space. It contains the molecular structures in a format that is readable by Turbo-Frodo.
When you start the program by typing>turbo, you will be prompted to give the name of the Heap File. If you answer with a non-existing Heap File name, the program will ask you whether you want to create this file or not. Answering yes will create an empty Heap File having the given name.
To fill the Heap File, you have to read the coordinate file about your molecule. This file may be in the Protein Data Bank format (PDB) or the Wayne Hendrickson format (WH). Accordingly, type turbo>read pdb or turbo>read wh.
The program will prompt you to give the name you wish to the molecule in the Heap File. This name should not be longer than 8 letters, otherwise only the first 8 letters will be taken into account.
The contents of the Heap File may be listed by typing turbo>list heap.
You first have to load the molecule
Moleculesdisplay by typing turbo>load "molecule", the "molecule" file name being that given in the Heap File.
Alternatively, you can load a molecule
Moleculesloaded by using its number in the Heap File. Once you have typed turbo>go, the molecule will appear on the screen in blue, which is the default color. At the same time as the molecule, the name of the program will also appear in the top-left part of the screen and XYZ axes in the top right, while the origin of the axes will be in the center of the screen. Remember that you can return to Textport by simultaneously pressing down the middle and right buttons on the mouse.
Since your molecule is loaded on the basis of its coordinates, it may happen that it does not appear on the screen. To center it, use the Textport option turbo>acon "residue number". This will put the alpha carbon (default atom) of the chosen residue in the molecule sequence at the center of the screen.
You may also find that only part of the molecule has been displayed. This occurs if the molecule was previously saved with the SPHERE option still active. To restore the display of the entire molecule, disactivate the display SPHERE option in the View menu then drag the mouse up the screen and from left to right. The following menu heads will appear: Help, Main, Color, Geometry, Editing, Modeling, View, plot, Tools, Surfaces, Molecules. (You can refer to Display Options for a systematic description of these menus.)
Dragging the mouse on the Molecule menu will make the name(s) of the loaded molecule(s) appear. The highlighted name(s) corresponds to the displayed molecule(s). If you want to turn off one of these molecules, click on the corresponding name, and the molecule will shadow. Click again on the same name to bring back the molecule.
You can change the color of the molecule by choosing the SAME option in the Color menu. If you have more than one molecule on the screen, you should have previously selected a specific one before by clicking on any of its atoms. A set of colors will appear, from which you can choose by clicking on one. This step may be useful if you want to discriminate between several displayed molecules.
A single-color display of the molecule is certainly not the most descriptive possible. You may therefore decide to color the molecule with various color codes, depending on the atom residue, side chain (in which case, the backbone will keep its original color), hydrophobicity, or on the secondary structure.
Another way of coloring the molecule is to discriminate between different streches by using the ZONE option in the Color menu. First, however, you must set up the stretches by using the following Textport option, turbo>zone. The program will prompt you to give the residue ranges. Answer accordingly, e.g., in the case of a molecule having 64 residues:
Residue range: 1 27, 28 56, 57 64
Residue range: 1 27, 27 56, 56 64
In the above example, there are three stretches running from 1 to 27, 28 to 56, and 57 to 64. To color these stretches, activate the ZONE option and successively select the colors chosen. The molecule will be displayed once the last color stretch has been set. Note that in the case of a CA representation, this setting will create gaps between alpha carbons 27 and 28, and 56 and 57. To overcome this problem specify the zones as follows:
Residue range: 1 27, 27, 56, 56, 64
You can decide to simplify the display by turning off all side chains. This procedure can be accomplished by activating the BACKBONE option within the View menu, thus displaying only the alpha carbons (the CA option within the same menu).
Once the CA display is on screen, you may decide to show only some of the side chains, for example to illustrate the active site on an enzyme. This step can be performed with the CA-SIDE-NAME option in the View menu. When this option has been activated, the program prompts you to give the names of the side chains you want to display. You can also display one or several types of side chains with the CA-SIDE-RES option in the View menu. When this option is activated, a list of side chain types appears. Click on the ones you want to display and validate your choice by clicking on "end".
You may also decide to display only part of the molecule. With the SPHERE option you can display a region centered on the most recently picked atom, corresponding to a sphere the radius of which can be changed using the following Textport command: For example, suppose you want a 5 radius sphere,
turbo>radius 5
The sphere will shift each time you center a new atom. (To center, simultaneously press down the left and middle buttons on the mouse.)
The last possible way of simplifying the displayed structure is to look at slices. The thickness and Z-screen coordinate of this section can be set up with the dials (or pseudo-dials). The scale can be displayed by using the Textport command turbo>slab-display, which helps you to localize this section in the space.
You may then decide to have a closer look at your structure. You can obtain various stereo views depending on what equipment you have. The stereo options are found in the Main menu.
Once you have chosen the visual representation you prefer, or diagram, you can animate it. If you have a button box connected to your workstation, you can rotate your molecule along the three axes as well as scale it. If not, you must activate the pseudo-dials box by using Textport's PSEUDO-DIALS option.
Warning: Before making any changes to a structure stocked in the Heap File, it is recommanded to duplicate it, so that you will eventually have one copy with the changes (which you will also save) as well as the original. This duplication can be accomplished by using the Textport option turbo>duplicate molecule. The program will prompt you to give the input and output file names.
The most recently picked molecule (which should not be larger than 1000 residues) will turn green, and additional pseudo-dial menus will appear in the lower-right corner of the screen.
With these pseudo-dials, you can rotate and shift the green structure along all three axes. This step can be done quickly and easily in the stereo mode, but if you are not keen on this mode, you can obtain an orthogonal view by clicking on ORTHO VIEW within the Main menu. This view will appear in an area in the top-right corner of the screen. If you are pleased with the result, pick yes to confirm the modifications otherwise pick no. If you pick yes, the changes will be saved on memory only.
If nothing further is done to save them, these changes will be lost when the session comes to an end. To save these modification in the Heap File, use the SAVE option in the Main menu or the SAVE ALL option if you have made changes to more than one loaded molecule. Do not forget that this will overwrite the entry, so be careful.
You can then transfer the result to a PDB, a WH, or an XYZ file by typing the following Textport options: turbo>make pdb or turbo>make wh or turbo>make xyz, as required. The program will prompt you to give the name of the output file.
You may want to keep a hard copy of your work. Check the position that the structure will have on the paper sheet by activating the PLOT FRAME option within the Plot menu. This will display the plotting frame, whose position can be adjusted by selecting its bottom-left corner with the left button on the mouse. Then the PLOT option within the same menu will prepare a Postcript file. The thickness of the line on the plot will be correlated with the color on the screen: the higher a color is on the color map, the thicker is the line.
There are a number of attractive representations, useful for example in a slide session for demonstrating how remarkable you molecule is. These representations are accessible for the selected molecule by means of the COMPACT SELE option in the View menu. Once you have chosen the representation you want, execute it by clicking on COMPACT DO-IT within the same menu. You can interactively handle a few of the representations, such as "Small Circle". For others, orientate your molecule before implementing the compact view.
You have just purified and sequenced a brand new protein or isolated and sequenced the gene involved. Comparisons between the protein's primary amino acid structure and those available from the protein data banks have shown the existence of a fairly good sequence homology with another protein, whose three-dimensional structure has been previously determinated and whose coordinates are available. You may therefore be anxious to see how singular your protein's structure actually is compared to the other one, and what structural differences between them might explain, for example, the functional differences observed.
The first thing to do here is to use a structure with known coordinates to model your protein. First, align the two sequences, either manually or with your favorite software, as carefully as possible, as they should not be merged with Turbo-Frodo. Then put the known structure on the screen by entering the following command: turbo>read pdb or turbo>read wh, as required.
The name of the molecule in the Heap File should be different from that of the coordinate file to avoid any confusion after changes are made in the structure. Then load this molecule by typing turbo>load "molecule", the "molecule" filename being that given in the Heap File. Once you have typed turbo>go, the molecule will appear on the screen, in blue, which is the default color.
You must then modify the sequence of the loaded molecule to fit the one you want to model. Simplify the displayed molecule by choosing the CA option in the View menu, which will display only alpha carbons, or BACKBONE, which will turn off the side chains. To correct the sequence, click on the alpha carbon of the residue you want to change, then go to the Editing menu and choose the REPLACE option. A window will appear allowing you to choose the new residue to be substituted. Move the mouse within the window until the chosen side chain has become yellow, then click on it and accept it with the previous dihedral angle values.
If the new side chain is longer than the former one, it will be replaced on the basis of the dihedral angles of the previous side chain in an extended geometry, without taking the neighboring into account. You will therefore have to correct this geometry manually in order to reduce as far as possible poor interatomic contacts. This correction can be achieved with the TORSION-AUTO option of the TORSION submenu found in the Modeling menu. Pick the residue you want to correct and you will see all the existing torsion angles displayed on the residue. A pseudo-dial will appear numbered according to the defined torsion angles. Using these pseudo-dials, you can now rotate the angles.
During this procedure, you can check the Van der Waals distance violations by activating the NEIGHBOR option in the Geometry menu, the cutoff being set to 1.8 , for example, using the Textport command RAD NEIGHBOR. The contact between the most recently clicked-on atom and any of its neighbors will be displayed until the distance between them has become larger than this cutoff.
A less trivial and often necessary procedure is that required to insert or delete a fragment. One or several residues can be deleted by selecting a residue and then using the DELETE option in the Editing menu. The program will ask you to confirm the deletion. Once you have deleted the loop, reconnect the disrupted backbone by using the BOND option in the Editing menu. Once this option has been activated, the cursor will be replaced by <1, which will prompt you to select the first atom in the new bond, and then by <2 , prompting you for the second atom.
Unfortunately, this brand new bond does not have the right geometry. Its geometry can be corrected, however, by using the REFINE option found in the Modeling menu. Once the option has been activated, the cursor will be replaced by <1, which prompts you to select the first residue in the strand to be corrected, and then by <2 for the last one, then start the refining procedure. The corresponding change of geometry during 20 cycles of refinement will occur on line on the screen. At the end of this procedure, you can either continue for the second refinement run (confirm the change by clicking yes) or go back to the former structure by clicking no.
On the other hand, you may want to insert a loop somewhere. First, disconnect the two residues between which the insertion will occur by using the BREAK BOND option and then start to insert the new sequence in sequential order by selecting the free carbon atom and then activating the INSERT option in the Editing menu. Select among all the residues the one you want. You will then be prompted to give the name of the residue. One way of naming it is to use the position of the insertion with a suffix: For example, if you want to insert 4 residues between residues 14 and 15, the additional residues will be named 14a, 14b, 14c, and 14d.
There are two possible insertion modes: the extended and helical fold modes. It is possible to switch from one mode to the other by cliking on the SHEET/HELIX option within the Editing menu. When you have inserted all the residues you need, connect the most recent residue to the free nitrogen atom resulting from the previous bond break. Use the REFINE procedure as described above to correct the geometry.
At the end of this correction procedure, and if you are satisfied with the outcome, use the SAVE option in the Main menu: This will ensure that the modified structure is stored in the Heap File. If you have made any insertions or deletions, change the numbering of the amino acids in your protein by means of the RENAME Textport option in order for these changes to be applied to the structure in the Heap File. You should then save the modified structure on disk as a .pdb or .wh file, using the make pdb or make wh Textport command.
Imagine you want to know the structure of a serine-protease inhibitor, i.e.,diethyl paranitrophenyl phosphate. Turbo-Frodo can design this molecule. In order to do so, start the program, selecting your current Heap file. You will then find all the appropriate tools needed to design the molecule de novo.
In the display Editing menu, select OPEN MODEL. The program will prompt you to give the name of your own dictionary. This dictionary is a file where small molecules may be saved. You will see lated the interest of such a file. If you do not have a dictionary, give a name. After giving a name, the program will ask you if you want to create a dictionary or not. A new menu pops up in the lower-right corner of the screen. The CONSTRUCT item is the main menu for construction. COMMON, SUGAR, and AM-ACIDS are the subdirectories of the fragment dictionary.
Select COMMON. In the lower part of the screen, several windows will appear, each containing a fragment. These fragments are basic parts of the puzzle. On the left-hand side are pseudo-dials which can be used to move these fragments. You can screen the fragment dictionary by using the lift located below the windows. You will then be able to browse in this dictionary. If you use the right button on the mouse to select a fragment, the fragment's name will appear. Picking an atom of a fragment with the middle button will give you the name of the atom. SUGARS and AM-ACIDS work in exactly the same way.
Select "--->USER DICT".You will see the name of the subdirectories, if any, of your own dictionary. You can go back to the main dictionary by selecting "--->databank". Now, start designing the inhibitor. Any atom added will be saved in the Heap File.
Go back to the COMMON directory and find the phosphate group. Then select CONSTRUCT and ADD FRAG. The cursor will change into <1, prompting you to give a selected fragment. Select an atom from the phosphate group with the middle button on the mouse.
The program will ask you what name the new molecule should be given in the heap and will display the fragment in red. This fragment will then be treated as a residue by Turbo-Frodo, i.e., like any other molecule. Change the color of the molecule and find the hydrogen atom. The design will proceed by creating covalent bonds between two atoms, each carrying a hydrogen atom.
Select ADD FRAG. The cursor will wait for you to give the first hydrogen atom. Choose that from the phosphate group. Once the >2 cursor appears, pick a benzene hydrogen. The covalent bond is obtained. SAVE if you want to keep this new fragment in memory. Add the nitrate in the same way to the benzene.
The following step would be used to create an ester bond between an ethyl group and benzene. There is no hydrogen once on the phosphate, but you can add the ethyl groups together by using the ADD ATOM option within the CONSTRUCT menu. After the two ethyl groups have been added, the construction will be complete. Do not forget to save this new molecule, using the SAVE option.
In the CONSTRUCT submenu, you will find other items: DEL ATOM and REPLACE ATOM can be used to delete and replace the most recently picked atom. DEFINE RESIDUE makes it possible to reorganize the molecule under construction .
Use BREAK BOND between the two carbons of an ETHYL group: The second carbon to be picked is that belonging to the methyl group. Then select DEFINE RESIDUE. The program will prompt you to give the type of the residue. Enter meth. You will now have 6 residues, including a "meth". If you now delete this residue and add a hydrogen atom to the trivalent carbon group, you will obtain a methyl-ethyl-paranitrophenol phosphate. You can see that the new O-C-H angle is 141¡6, which cannot be correct in the case of a methyl group. The DEFINE ANGLE option makes it possible to enter the right value, i.e., 250¡5. You can correct a distance in the same way.
The brand new molecule can be used as a basic part for further constructions, provided this molecule has been inserted into your own dictionary: Select DATA BANK within the CONSTRUCT menu, then USER DICT. A window will appear, and you will then be able to play with your dictionary. You can list its contents, create subdirectories, and so on. Select INSERT molecules with the right button on the mouse. If your dictionary is new, you will have to create a directory by using CREATE DIRECTORY.
Go back to the display by using END and use "---> USER DICT" to display your own dictionary. You can see that your molecule is now being treated like a fragment in its own right.
CHAPTER 17
X-Ray Users' Tutorial
This chapter describes how a Turbo-Frodo session may be run after obtaining an electron-density map
Electron-density map.
Turbo-Frodo uses the classical DSN6 maps from the older Frodo packages. These maps can be calculated from Fourrier maps with the Mappage program, which can read the specific Fourrier map format. Some of the most classical crystallographic packages are CCP4, PROTEIN, and XPLOR.
A Mappage program for Silicon Graphics equipment is provided along with Turbo. This program is located under the $TURBO_DIR/map subdirectory.
Whether your map has been produced with XPLOR or CCP4, you should be able to change its format with the supplied Mappage program. For the XPLOR package, there are no particular instructions. For the CCP4 package, however, you should scale the density with the inverse of the rms deviation from mean density. You can see these instructions while running the Mappage program. The output map is the file used by Turbo-Frodo.
Since maps are logically linked to a molecule, you need to have one molecule in the heap, loaded in memory. You can load up to five maps for each molecular entry:
You can redefine the map parameters either when loading or afterward from the pull-down MAPS submenu. The default levels are 1 for the first map (usually a 2fo-fc) and 2 for the four others (fo-fc).
CHAPTER 18
NMR Users' Tutorial
This chapter describes how a Turbo-Frodo session may be run after calculating a protein structure using a distance geometry program (DISMAN
DISMAN, for example) or a constrained molecular dynamics program (XPLOR, for example). These programs, using as input a set of distance constraints based on NOESY spectra, or other sources such as the disulfide bonding or hydrogen bonding, have produced an output file containing the coordinates of the molecule studied.
Displaying a Calculated Structure
You first have to check the format of this output file, which has to be under a PDB or WH format. Since this structure is based on NMR data, you do not have to worry about hydrogen generations, which should be present in the coordinate file. Turbo-Frodo can also work with the necessary pseudo-atoms if no stereospecific assignment has been carried out.
The format of the input file containing the distance constraints also has to be written in the format used by XPLOR. Otherwise, you will have to make the corresponding corrections everytime (see the Necessary Files chapter about correct format). You then have to fill up a new or previously existing Heap File with your new structure by using the following commands, turbo>read pdb or turbo>read wh.
You should then load this molecule by typing turbo>load "molecule", the "molecule" file name being that in the Heap File. Alternatively, you can load a molecule by using the number it has in the Heap File.
Evaluating the Calculated Structure's Quality
To make the best use of the program, you should also load the distance constraints file used for structure calculation or refinement. This step can be achieved by typing turbo>load nmr. The program will ask you to give the name of this file. Answer accordingly, assuming this file to have the right format (see the Necessary Files chapter about the correct format). The program will then list the atoms used in the distance constraints file but not in the PDB file that you have used. This problem arises mainly when the names of pseudo-atoms that you have used are not correct according to the XPLOR format.
You can now look at your new structure together with the constraints you have used to calculate it by typing turbo>go. The structure will appear in light blue together with the constraints, which will be colored depending on the violations, if any:
Green: The distance between the two protons (or pseudo-atoms, if any) is lower than (or equal to) the constraints used.
Orange: The constraints used is violated by not more than 0.2 .
Red: The constraints used is violated by more than 0.2 .
Note: This tolerance of 0. 2 can be modified with the TOLERANCE Textport option.
The constraints can be turned off by dragging the mouse through the Tools menu, until NMR is highlighted. If you click on the right button, the NMR submenu will appear. You can then activate the ON/OFF option, which will alternatively turn the NMR constraints display on and off.
Alternatively, you can turn off the molecule display by dragging the mouse to the menu on the very left and clicking (by using the right button) on the highlighted molecule name (the one that is currently displayed). This will make the structure vanish while the distance constraints are still displayed.
Since a standard distance constraints input may contain up to 10 constraints per residue, displaying all of them at once often gives a messy picture. To simplify matters, you can display only the violated constraints by choosing the VIOLATION option in the NMR submenu.
Another way of estimating how the structure was calculated is to draw a picture of the constraints over the molecule sequence. This step can be accomplished by choosing the COL REST RES and COL REST ATO options, which color the residues or atoms, respectively, depending on the number of violations in which they are involved. A yellow residue is an unconstrained one, and the colors range from red (few constraints) to blue (many constraints). With this option, you will immediately see which parts of the molecules have been poorly determined and how flexible they are. This may be roughly compared to the B-factor in X-ray crystallography.
You can also carry out this estimation on one particular residue, using the DISPLAY RES option in the NMR submenu. This will cause the constraints relating to the clicked-on residue to be displayed.
To finally estimate the quality of this initial structure, you have to check its geometrical accuracy by calculating the phi and psi conformational angles. This can be done by picking the RAMA ALL option in the Plot menu.
Correcting the Calculated Structure
Now that you have estimated the quality of the calculated structure in terms of violations, and the geometry of its residues, you can try to correct it manually. As long as there are not many violated distance constraints (otherwise it would be better to sort out the input file constraints first), you can try to satisfy them by modifying the structure while displaying the constraints.
After displaying the constraints (see above), change the geometry of the side chain involved using the TORSION-PICK option in the Geometry menu. You will first have to define the torsion angles by picking successive atoms. Note that four atoms are needed to define the first torsion angle, and then for each additional picked atom, a new angle will be defined. Up to six torsion angles can be defined, and these will be numbered on the displayed structure. Validate your choice by clicking on yes, and the pseudo-dial will appear, corresponding to the defined torsion angles. Using these pseudo-dials, you can rotate these angles. The displayed constraints will follow, and their color will change depending on the distance. You should try to keep the modifications to a minimum, and stop rotating, for example, when the constraints' color turns green.
During this procedure, you can check the Van der Waals distance violation by activating the NEIGHBOR option in the Geometry menu, after setting the cutoff to 1.8 , for example, by using the RAD NEIGHBOR Textport command. The contact between the most recently picked atom and any of its neighbors will be displayed once the distance between them has become smaller than this cutoff.
At the end of this correction procedure, if you are satisfied with the outcome, confirm by picking yes or go back to the original structure by picking no. This will save the changes made during the open Turbo-Frodo session only. If you want to store the result on disk, activate the SAVE option in the Main menu. This will save the modified structure in your Heap File.
You can now submit the corrrected structure to whatever energy minimization or simulated annealing procedure you want. However, you will first have to produce the corresponding coordinate file, by using the following Textport command, turbo>make pdb or turbo>make xpdb, if you want to refine the structure with XPLOR.
Alternatively, Turbo-Frodo makes it possible to calculate a list of theorical nOe peaks, provided you have the list of protons in your protein. At present, this is still only a rough procedure, consisting of extracting couples of protons separated by 5 or less and calculating a peak volume without taking any relaxation matrixes into account. However, you can load a list of the peak on your spectrum, either manually or automatically, provided you use the EASY program as the assignment software. You will thus be able to label previously unassigned NOESY peaks.
For this purpose, use the following Textport command, turbo>nmr-spect. The program will prompt you to give the input proton list, which has to be written in the EASY format or directly produced by it (see the Necessary Files chapter for an example of this format).
The output files are a list peak and a .ref file, which have not yet been used. The list contains a list of peaks for each peak from left to right: the number of the peak, the chemical shift of the first proton, the chemical shift of the second proton, the color of the nOe peak (if displayed using EASY), and a roughly calculated volume (without taking any relaxation matrixes into account). The other entries have not yet been dedicated.
CHAPTER 19
Playing with Surfaces Tutorial
Turbo-Frodo offers several ways of displaying the surface of your protein.
Once you have loaded the molecule, you can calculate the Van der Waals volumes of all or part(s) of the molecule. This procedure can be accomplished by activating the ADD VDW SURF option in the Surface menu. The cursor arrow will be replaced by <1, prompting you to give the first residue, in the stretch of which you want to calculate the surface. After this selection, <2 will appear, prompting you to give the last residue in the stretch. During the calculation of surface, the Textport window will appear, giving you information about what is happening. Note that the Van der Waals surface is calculated for all the heavy atoms in the stretch, no matter which mode is used for the molecule. This surface can be colored differently by activating the COLOR VDW option in the Surface menu, and then choosing the color within the Color menu.
You can add another Van der Waals surface by choosing the ADD VDW SURF in the Surface menu, or start another calculation by activating NEW VDW SURF. If you have several calculated surfaces on screen, you can display one or several of them by selecting them in VDW MENU. Those that are selected will be highlighted in yellow.
You can choose the number of dots per square angstrom via the following Textport option, turbo>vdw-density, and answer with what follows:
This is the surface of a molecule accessible to a water molecule. It can be calculated with the ms program in TURBO-DIR/ms, after updating the ms.in and ms.com files. Copy the ms.in and ms.com files in your working directory before updating them.
Once the molecule that needs to be calculated has been loaded, go to the Textport window and use the >make ms command. The program will then prompt you to name the output file. Give the output file a name with the .atm. extension. This file is the coordinate input file of the ms program.
Then, outside Turbo, for example by opening another window, edit the ms.com file, which is in the TURBO-DIR/ms subdirectory and substitute your_molecule.srf file with the name of your molecule, keeping the extension as it is. The your_molecule.srf file is the output surface file.
Edit the ms.in file and change the coordinate input file name your-molecule.atm into the name previously given as the output surface file. Then run the ms program:>ms.com. Since this is fairly time consuming, it is best to run this program as a batch job.
The output surface file of ms.com has the .srf extension and will be used if you type the turbo>connolly Textport command. You will be asked whether you want a simple calculation or to link the points to atoms. In the first case, you do not need to have a loaded molecule, but this will produce a single object, and Turbo will be unable to color it depending on the residue type or the side chain.
This will prepare the real surface file having the .ts extension, which can be displayed by typing turbo<load surface. You will then be asked to name the surface file (the one with the .ts extension). The program knows what kind of surface file is being used. If the surface is to be calculated including the atom dependence, you must have previously loaded the molecule.
This is the surface calculated from the Connolly surface which is displayed in the form of lines obtained by interpolating the Connolly surface points.
Using the Connolly Surface File as above (the one obtained with ms.com, having the .srf extensions), run the prep_spl programs followed by make_spl, which are to be found in the TURBO-DIR/ms directory.
>prep_spl
You will be asked to give the thickness of the slice and the name of the output file which is used as the input file for the program.
>make_spl
Once again, give the name of the output file thus created. Once you have obtained this file, use the following Textport command to read the Spline Surface File turbo>spline. The program will then produce an output file with the .ts extension. To display this file, use the Textport command turbo>load surface.
Warning: If you calculate both the Connolly and the Spline surfaces of the same molecule, you will produce the "your_molecule.srf" and "your_molecule.spl" files. In this case, the program will transform both files into "your _molecule.ts", thus overwriting the first file to be produced. To overcome this problem, give the two surface files different names.
These surfaces can be colored differently by activating the COLOR SURF option of the Surface menu, then choosing the color from the Color menu. You can switch off the surface(s) by activating the NO SURF option within this same menu. Assuming you have calculated and loaded several surfaces, you can display one or several of them with the SURF MENU option, by selecting the required surface(s).
CHAPTER 20
Non-Crystallographic Symmetry Tutorial
Non-Crystallographic Symmetry (NCS) in Turbo-Frodo relies upon four successive steps: definition of the domains related by the symmetry, definition of the symmetry operation, display of the NCS-related molecules, and display of the NCS-related maps. These different steps are managed by the Textport options DOMAIN and NCS, respectively, and by the displayed NCS submenu.
Definition of the Related Domains
You must first be able to define the different domains of the molecule. A domain can be one entire molecule if you have several in the asymmetric unit. (In this case, all molecules of the asymmetric unit must be in the same heap entry.) A domain can also be a part of a molecule if this molecule has an internal symmetry. You can define a domain by using the Textport option DOMAIN. One domain can include several zones and is automatically saved in the heap.
Definition of the Symmetry Operation
Turbo-Frodo considers pairs of domains, and thus one complete symmetry operation (rotation matrix and translation) between two domains. If you have three related domains, you have to define the three related pairs (two may be enough, depending on how you want to work).
You have three different ways of indicating the related pair and its symmetry operation: You can either read it from a file, read it from another heap entry, or calculate it. All of these options are managed by the Textport NCS handling.
In case you have to calculate the rotation and translation, choose the following sentence, 1-Define an operation symmetry, in the NCS handling. This calculation works as the RIGID option does: It is a rigid body superposition of the two domains. You must therefore indicate some of the atoms from the two domains, which fit together. If you need help or more information on the rigid body superposition, you can refer to the RIGID help. Once defined, the pair and symmetry operation are saved in the heap, so you do not need to save all over again. The inverse operation is also calculated.
If you have a new heap entry, but NCS has already been defined for another one, you can just copy NCS from one molecule to the other. In order to do so, choose the following sentence, 4-Read a matrix from another heap entry, in the NCS handling. (The source molecule has to be loaded.) For this option, both domain definitions and symmetry operations are copied. You can also specify a name file that contains the symmetry operations (3-Read a matrix file). This file must have the same format as the one dumped by the display option PRINT NCS. It is a way of copying the symmetry from one molecule to another but not the domains. The domains of both molecules can be different.
Display of the NCS-Related Molecules
When domains and symmetry operations are defined, you can work with the Non-Crystallographic Symmetry. The domains can be separately colored by using the numerous options in the Color menu. To display the related molecule, you must use one of the two options of the NCS submenu (Tools menu): APPLY MATRIX or AUTO APPLY.
For APPLY MATRIX, the different related pairs are proposed, and you have to select one or several pairs. You can note that the inverse operations are also proposed. On display, both domains of a related pair are superimposed. This is only a display operation that does not affect the coordinates of the transferred domain. (A mistaken SAVE will not have any consequences.)
If you choose the AUTO APPLY option, all the domains related to the centered one (CONTOUR residue) will be transferred. The symmetry applied may change if you CONTOUR on another residue (belonging to another domain).
Display of the NCS-Related Maps
You may also be able to apply the symmetry to the displayed maps. You can display a map box of a domain with the NCS MAPS submenu. The new box has at first the same parameters as the original map: same levels and colors. These parameters can be changed by choosing the OP knot of the NCS MAP menu. You can ask for different levels and colors, dashed lines or solid ones, but the box width remains unchanged, as does the axis contour.
You will also be asked for the type of map that you would like to be displayed. Indeed, not only the map difference can be displayed, but also the average density. The map difference is represented by the absolute value of the difference; furthermore, only positive values are considered. The average density represents the average between all symmetric boxes and the original one. These operations are applied between the original map box and the transferred ones. If you work with a three-fold axis, the two transferred map boxes will be superimposed to the original map box, except if you choose average density, where you see only one superimposed average map.
CHAPTER 21
Ab Initio Building Tutorial
Ab initio building is a new and main feature of Turbo-Frodo, and from now on, you will be able to work with an MIR map. We propose here a new way of building a protein de novo, with the presentation of a new tool: the Trigonal Planar Pseudo Residues
Trigonal Planar Pseudo Residues (TPPRs).
While moving along your new MIR electron density map, you will see a stretch of helix, or B strand. You can place C Alpha atoms in a very easy way by adding some TPPRs anywhere you wish on the density map. By asking to connect these TPPRs, you will spot the CB atoms and will be able to describe the side-chain density. When asking to align those fragments with your protein sequence, you will orientate the fragment and place the N and C atoms. Finally, a special algorithm of replacement will fit the O carbonyl and the whole peptidic bond in the density map. All these features can be found under the displayed BUILDING and BONES submenus. The few required parameters may be set via the Textport options.
You are now able to move along an empty density map. As you contour an atom, press down the left and middle mouse buttons anywhere on the map, and the whole image will center on the point that you have selected. When selecting a point on the screeen, you in fact select a line that is orthogonal to the screeen and to the displayed slab length. The new center is the highest density point found along this line. When several maps are displayed, a research is carried out among all the maps. If you do not get the expected results, you can decrease the slab, which may cause the selection to become more accurate.
Adding New Entities in the Density Map
To build ab initio in the electron density map, the strategic choice relies on the idea of TPPRs. A TPPR is constituted of four atoms: CA, C1, C2, and C3. The CA atom is in the center of the trigonal plan formed by the three other atoms. The aim is to find the best position for the CA atom, and in order to do so, we chose to handle a TPPR that is easy to place in the crossings of the density that you can spot. This is a planar residue because when looking at an empty density map, evidence is not always available to orientate the residue (or to decide which atom is N and which one is C) and therefore to orientate a real tetragonal carbon. When you click in the density, the position of the TPPR is defined with the same algorithm as that used when centering in the density map. Therefore, a Real Space Refinement of the only orientation places the three branches of the TPPR in the density map.
Localization of the TPPR center Real Space Refinement of
on the highest density point the orientation to determine
the best density fit
Sometimes it can be useful to refine the center position, and therefore a Real Space Refinement that performs translations can also be added to the previous one.
By selecting the ADD TPPR option of the BUILDING submenu, you will be able to add a TPPR on the map. Select one of the following two options, Build Mode 1 (BM1) or Build Mode 2 (BM2), at the right of the screen, then select a point in the density by clicking on the left mouse button. The first option (BM1) allows an RSR of the TPPR position by translation, while the BM2 option does not. BM1 and BM2 are useful, depending on the shape of the electron density. When several maps are displayed, you must choose the working map with Textport's WORK-MAP option in order for RSR to be performed on this particular map.
For the first TPPR added, the program will prompt you for a residue name, suffix included (e.g., 24A). The following residue names will be automatic increments of this first one (e.g., 25A, 26A, and so on). The TPPRs are created with a TPPR residue type, which can be changed when activating the CA LABELS option of the BUILDING submenu. A label window appears at the lower-left corner of the screen. If you select the residue type, you may be able to choose another type in the appearing submenu. Residue types include Short, Medium, Long, and Aromatic and can serve as a way to describe the shape of the density, since it is difficult to recognize the residue type in an MIR map. If you select the comment item of this label window, you can attach a comment to your residue. The comments can be listed or dumped in an ascii file by using the following Textport orders, LIST COMMENTS and MAKE COMMENTS, respectively.
Once you have added some pseudo residues in the density, you have to connect them. While working, you may have added the pseudo residues in a random order, and when you list all the residues or ask for a CA trace, the residues do not have the good connectivity. You can create this connectivity with the MAKE FRAGS option of the BUILDING submenu. This step will connect the CA atoms that are 5 angstroms apart, but this cutoff distance can be changed by using the DIST-CA Textport option. Furthermore, the consistency of the connectivity will be checked. A bond that crosses the density (that is, goes through an empty area) does not exist, although the distance can be below the cutoff. For each CA, the program will choose the two bonds whose density fittings are the highest among the allowed neighbors (distance smaller than the cutoff). The density cutoff is set to the first displayed level of the working map.
When proceeding in this manner, a distance cutoff set between 4.5 and 5.5 is recommended but is also highly dependent on the electron density map that you have. Sometimes, for example, you will find that the distance cutoff must be diminished when connecting residues in helix conformation, otherwise the two neighbors chosen for a given CA will not be the two expected ones.
Identifying the Pseudo Residues
The fragments that you have just assembled are neither orientated nor placed along the target sequence. This target sequence (that is, the currently built protein sequence) must be loaded with Textport's LOAD SEQUENCE order, and it will be attached to the molecular entry and automatically saved in the heap file. Go through the alignment screen by activating the ALIGN FRAGS option of the BUILDING submenu. You will discover in the upper part of the screen the fragments that you established before, and in the lower part the target sequence. (Pressing down the left mouse button will end the alignment screen and return you to the molecule).
When selecting a fragment with either the right or middle mouse button, the alignment is performed, and the ten best solutions are proposed: Just choose one of the ten buttons in the upper-right corner of the screen. The fragment is aligned under the displayed target sequence with a color code: green (10-point score) for correct matching, red (0 points) for bad matching, and yellow (5-point score) if both residue types belong to the same group (Short, Medium, Long, or Aromatic). For each proposed solution, the score (or sum of residue scores) is given, as well as the orientation of the fragment. You therefore realize that the residue type that you gave to the TPPR is extremely important and that the shape of the density must be well described.
If you are satistied with a proposed solution, you can confirm it with the following consequences: The fragment will disappear from the pool of fragments, the pseudo residues will be renamed after $x (x depending on the place the pseudo residues now have in the target sequence), and finally the atoms will be named after C, N, and CB as the fragment is now orientated. Overlapping is managed and you can decide to put a fragment back in the pool if you confirm another one at its place.
You now need to replace the connected TPPRs with real residues, and this according to the target sequence. The MULTI REPLACE option of the BUILDING submenu will perform this replacement and will replace as many pseudo residues as you would like. Simply select the first and last TPPRs of a fragment. If you wish, all the pseudo residues will be replaced by a real residue according to the target sequence. If you do no wish to do so, answer no to the appearing question; the residues whose type does not correspond to the target sequence (e.g., Short, Aromatic, or anything else) will not be substituted.
The MULTI REPLACE option is not as simple as the REPLACE RES option of the Modeling menu. You noted that up to nown, the O carbonyl (which is as important as the side chain) is missing. This option is in fact an option to place the O, and therefore the whole peptidic bond, but according to the density map. An imaginary peptidic plan (CAi, Ci, Oi, Ni+1, CAi+1) is rotated on the CAi-CAi+1 axis of two TPPRs, and the best density fit is kept. This implies that Ci of the TPPR(i) is moved, as well as the Ni+1 of the TPPR(i+1),to build a real peptidic plan with the best density fit. The CB of the former TPPR(i) is not moved and is the starting point for the side chain which comes in the most frequent conformation, according to the rotamer library of J.W. Ponder and F.M. Richards. It goes on building peptidic plans and side chains until the last residue you asked for. For this last residue, the oxygen is placed without the help of a map.
Once the completed fragment is built, 20 cycles of refinement are performed (with the REFI algorithm) to correct the geometry. Note that this option does not work only with the TPPR, but is also quite performant to correct a peptidic bond between 2 residues, even in the case of the Molecular Replacement method.
A toolbox has been set to adjust the conformation of the residue according to the density map. A permanent menu appears when activating the TOOLS-BOX option of the BUILDING submenu. To find the best density fit manually, you can use the RESI ROT/TRANS option, which is an adaptation of the FRAG ROT/TRANS option. The whole residue is automatically selected for both translation or rotation, and there is no need to break some bonds to isolate the residue. The rotation center is the last picked atom on that residue. The ROTAMERS option proposes all the different rotamers on the side chain, as they have been defined by J.W. Ponder and F.M. Richards, and offers a way of getting the conformer that is best suited to your density. The RSR MAIN and RSR SIDE options perform Real Space Refinement of the residue's main chain (CB included) and side chain, respectively. All these options work on the most recently picked residue.
You may wish to get help from a bones map. To obtain the bones map, you need to run the following supplied program, $TURBO_DIR/map/bones_turbo, and then load this new bones map by using Textport's LOAD BONES order. When selecting a bones point, you will see a label window that appears in the bottom-left corner of the screen. In this label window, you may change the name (or label), the bones type, or the residue type of the last picked bones point by clicking with the right mouse button on the appropriate place. The bones type is one of the four following types: C ALPHA, MAIN, SIDE, UNDEFINED; choose one in the appearing pop-up menu. You will be able to change the residue type for those that have a C ALPHA bones type. These residue types include Short, Medium, Long, and Aromatic. When moving along the bones map, and with the help of the electron density map, you should spot all the bones points which seem to be at a C Alpha place and identify them properly. You can then have a sort of CA trace: A C Alpha bones trace, by selecting the BONES C ALPHA option of the BONES submenu.
Once you have located these CA points, you may want to have real residues. You can create residues of only one CA atom by activating the CREATE CA option of the BONES submenu. Select the first and last CA bones points of a connected set. The new residues will come at the exact location of the identified C Alpha bones point and will also have their residue type. Above all, however, you can create TPPRs by activating the CREATE TPPR option of the BONES submenu. These TPPRs will be created via the BUILD MODE 1 (both rotation and translation performed), where the C Alpha bones point will be the starting point (first location of the CA atom). When you have the TPPRs, you can enter the alignment screen and thus identify the fragment that you have built.